OpenAI and Broadcom Unveil Jalapeno AI Chip for LLM Inference on June 25

According to OpenAI, the company and Broadcom unveiled Jalapeno on June 25, a custom AI accelerator designed specifically for large language model inference. Developed in partnership with Broadcom and Celestica, Jalapeno represents the first component of a planned multi-generation compute platform aimed at improving speed, efficiency, and accessibility of advanced AI systems. The chip was built from internal research into LLM inference requirements and incorporates kernel optimization, memory handling, networking, and serving systems. Early engineering samples are already running machine learning workloads in laboratory environments, including those associated with advanced models such as GPT-5.3-Codex-Spark, operating at target frequency and power levels. The architecture emphasizes reduced data movement and balanced resource distribution across compute, memory, and networking, designed to work across different large language models.
Disclaimer: The information on this page may come from third-party sources and is for reference only. It does not represent the views or opinions of Gate and does not constitute any financial, investment, or legal advice. Virtual asset trading involves high risk. Please do not rely solely on the information on this page when making decisions. For details, see the Disclaimer.
Comment
0/400
No comments